Fig. 4.
Computation of the library of context profiles representing local sequence contexts. From a database (NR30) of 1.5M groups of aligned sequences covering the NR database, we select the 50,000 most diverse alignments and enrich these with homologs from a single BLAST search. The alignments are converted to sequence profiles and 1M profile windows are randomly sampled and used to train K context profiles (K = 500, 1,000, 2,000, 4,000) with the expectation maximization algorithm.
